Transcript Warm and Cold Ion Linac

Warm and Cold Ion Linac: Comparison
and Optimization
March 30, 2015
Content








Economics of NC and SC linacs
Peak and accelerating fields in NC and SC structures
Review of pulsed ion Linacs
Main difference between NC and SC structures in application to
multi-ion linacs
High-performance QWRs and HWRs
Cost-efficient design of the SC section: reduce number of cavities
and cryomodules
Lorentz detuning and its compensation
Conclusion
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
2
MEIC Requires an Multi-Ion Linac
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
3
What Do We Know about NC and SC Linacs?
 Linac cost includes 3 major components
–
–
–
–
Accelerating structure
RF system
Cryoplant and helium distribution system (SC linac)
Everything else (building, vacuum, water cooling, controls,
diagnostics,…)
 Normal Conducting Pulsed Linac
–
–
–
–
The most expensive component is the RF system
High power (many MWs) RF systems are cost efficient
Multi-gap, long accelerating structures are cost efficient
Accelerating gradients are limited by breakdowns which is known as
Kilpatrick limit
 Superconducting Pulsed Linac
– Accelerating structure is very expensive
– Cryoplant is considerable cost
– RF is low-cost especially for low power linacs
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
4
Peak Fields in Accelerating Cavities
Accelerating gradients are limited by peak fields



Normal conducting structures made
from copper
Kilpatrick limit was introduced in 1950s,
empirical formula
In modern pulsed structures, electric
field exceeds Kilpatrick limit by a factor
of ~2
8.5
f [ MHz ]  1.64 E e
2
K

EK
, EK [ MV / m]





100
100
Superconducting structures
Peak magnetic field is limited by
quench, theoretical value is ~200 mT
at 2K
Peak electric field is limited by the
surface quality. ~120 MV/m can be
achieved
Peak fields can not be measured
These ratios are known from the
simulations of the resonator design
80
EPEAK
,
E ACC
60
f ( Ep)
40

20
0
0
6
6
8
10
Ep
P.N. Ostroumov, Warm and Cold Ion Linacs
12
12

BPEAK
E ACC
EACC can be obtained experimentally
from the stored energy
EPEAK can reach ~60-70 MV/m in
pulsed oepration
Collaboration Meeting
March 30, 2015
5
Autophasing, Synchronous Phase in Multi-gap
Standing Wave Structure
Lc
1
1 0.8
0.6
E ( ωt)
0.4
0 0.2
0
2
1
 1.571
1
1 0.8
0.6
E ( ωt)
0.4
0 0.2
0
2
1
 1.571
 = 0, maximum acceleration
0
ωt
t  2 , t flight  Tperiod
1
2
1.544
Lc   averagecTperiod   average
 =-30, stable acceleration
0
ωt
1
2
1.544
QeVeff cos  s
W
W  QeVeff cos  s ,   1 
 1
AW0
AW0
Q is ion charge, A is ion mass number
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
6
Acceleration of Ion Beams
 Due to limited extraction voltage in ion sources, RF acceleration is
applied at very low velocities
 To have practical length of the acceleration period, the RF frequency
must be low
Lc   average    average

c
f
Transverse dimensions of accelerating structures are large, ~
1
f
 RFQ is the first accelerating structure after the ion source
 IH type structures are popular right after the RFQ
 Alvarez structure is used at GSI after the IH structure
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
7
RF Linacs
RF Linacs
CW
Pulsed
SC
NC*
ISAC-I
RIKEN inj.
LEDA RFQ
SARAF RFQ
ATLAS RFQ
*Low-energy,
several MeV/u
Heavy-ions
ATLAS
ISAC-II
INFN
SARAF
SPIRAL-2
FRIB
Project X
EURISOL
RAON
ADS projects
P.N. Ostroumov, Warm and Cold Ion Linacs
NC
SC
Protons, H
SNS
Synchrotron
CERN SPL
Injectors (FNAL,
ESS
KEK, CERN, IHEP….)
MMF (Moscow)
SNS
LANSCE
Heavy Ions
Synch. Injectors at GSI
LEIR - CERN
BNL injector
Carbon therapy
Collaboration Meeting
synchrotrons
March 30, 2015
8
CW Linac: NC or SC ?
 Required wall plug power to create accelerating field
Pw
2
Veff
LR sh
, Rsh  function( f ,  , aperture),   0.3
 Typical example: 1 GeV CW linac
E acc  3 MV / m,  s  25 , Rsh  30 M / m
L  368 m, and Pw  368 MW
 Superconducting CW linac is much more economic than NC
 Both pulsed or CW SC linacs require NC front end
– Multi-gap cavities are cost efficient due to fast change of velocity

~0.3 to 10 MeV/u depending on q/A and duty factor
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
9
Low Duty Cycle Pulsed Linacs, Normal Conducting
 For injection into synchrotrons we need higher  and  to control
space charge in the ring
 For acceleration of ions we need A/Q times higher total voltages
compared to protons to reach the same velocity
– Acceleration of ions with low Q and high A requires much more voltage
then for protons
 For pulsed linacs it is cost efficient to have multi-gap accelerating
structure and pulsed high-power RF system, several MW pulsed
power
 GSI Linac, UNILAC – 11.4 MeV/u


– RFQ (36 MHz), Q/A  4/238
– Alvarez (108 MHz), Q/A  28/238
Linac-3, CERN - 4.2 MeV/u, Q/A  27/208
– RFQ (100 MHz), IH structure (3 tanks), 100 MHz, 200 MHz
New BNL injector, 2 MeV/u, Q/A 1/6
– RFQ, IH structure (1 tank), 100 MHz
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
10
Drift Tube Linac (DTL) for Protons
 Protons, f=200 MHz, resonator diameter is ~1 meter
 From 0.75 to 200 MeV/u
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
11
J-PARC DTL
 F=325 MHz, H-minus, 3 MeV to 50 MeV
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
12
GSI Linac
 IH structure, 36 MHz
up to 1.4 MeV/u
P.N. Ostroumov, Warm and Cold Ion Linacs
Alvarez, 108 MHz
up to 11.4 MeV/u
Collaboration Meeting
March 30, 2015
13
Normal Conducting IH structure
 U. Ratzinger (Frankfurt) Group has built it for the BNL EBIS injector
project
Courtesy of J. Alessi (BNL)
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
14
Room Temperature Linac at BNL
 RFQ and IH structure

EBIS
4-rod RFQ
IH-structure
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
15
Fixed Velocity and Variable Velocity Structures
V sin( t   0 )
Normal Conducting
Beam
β/2
V sin( t   1 )
V sin( t   3 )
Normal or Super Conducting
V sin( t   2 )
Beam
β
/2
P.N. Ostroumov, Warm and
Cold Ion Linacs
OPT
Collaboration Meeting
March 30, 2015
16
Front End of Multi-Ion Linac
 For low duty cycle linac, NC front end up to ~5 MeV/u is cost
effective
 The cost of pulsed RFQ and IH structure is low: much less than the
cost of CW structures
 For this section of NC linac, the main cost is in RF system
 Much more cost-effective than SC structure due to large range of
velocity change
– Multi-gap SC structure – high cost
– Or many short SC cavities are required - expensive
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
17
Linac for MEIC: from Proton to Lead
 For heavy ions we need
RFQ
IH
Q/A=30
IS MEBT
13.2 MeV/u
Q/A=67
Stripper
 Assume, we need 100 MeV/u for lead ions
 Higher energy for light ions is preferable to control space charge in the
ring (light ion beams are more vulnerable to space charge)
 In low duty cycle linac NC front end is more cost-effective, ~5 MeV/u
 For protons we need effective voltage
VEFF 
100 MV
 115MV
cos  s
 For lead ions we need effective voltage
VEFF 

13.2  A
86.8  A
13.2  208 86.8  208



 417 MV
Q1 cos  s Q2  cos  s 30  cos  s 67  cos  s
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
18
Options
 NC pulsed Linac for lead ions up to 100 MeV/u
– We need 417 MV total voltage
– If linac is based on long multi-gap cavities it can be used for protons too
• The phase velocity is given by the geometry of the linac
• Lower the voltage and power to maintain synchronous motion of light particles
• Protons will gain total voltage 115 MV and can be accelerated up to 100 MeV
– If linac is based on 2-3 gap cavities like SC linac
• Protons can be accelerated to higher energies, ~280 MeV
• Peak fields are limited by breakdown voltage (Kilpatrick limit), this volatge is
lower for low frequencies
• Factor of 5 larger number of cavities as compared to SC option will be
necessary
 SC pulsed Linac
– Provides higher accelerating voltage per cavity than NC structures
• About factor of 4 higher accelerating voltages per cavity than for NC cavity
– Number of 2-gap cavities is 107 to accelerate lead ions from 5 MeV/u to
100 MeV/u
– Provides 100 MeV/u lead ions and 280 MeV/u protons
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
19
Basic Parameters of the Linac
 Linac layout
RFQ
IH
QWR
QWR
IS MEBT
Normal conducting
Stripper
Ion species
Superconducting
Ion species for the reference
design
Kinetic energy of H & Pb ions
Maximum pulse current
Light ions (A/q3)
Heavy ions (A/q>3)
440
430
Linac total voltage (MV)
HWR
420
Pulse repetition rate
Pulse length
Light ions (A/q3)
Heavy ions (A/q>3)
410
400
390
380
0
5
10
15
20
Stripping energy (MeV/u)
P.N. Ostroumov, Warm and Cold Ion Linacs
25
30
Maximum beam pulsed power
Fundamental frequency
Total
length
Collaboration Meeting
H to Pb
208Pb
MeV/u
280 & 100
mA
mA
2
0.5
Hz
up to 10
ms
ms
0.5
0.25
kW
MHz
m
680
115
110
March 30, 2015
20
Focusing
 Focusing with SC solenoids is cost-efficient
 Return coils are used to minimize fringe fields, no iron or mu-metal
shielding is required
9-Tesla SC solenoids in helium vessel
ATLAS Cryomodule
P.N. Ostroumov, Warm and Cold Ion Linacs
6-Tesla SC solenoids in helium vessel
PXIE Cryomodule
Collaboration Meeting
March 30, 2015
21
Quarter Wave Resonators
 This data for CW mode, even higher fields are possible for pulsed
mode
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
22
HWR, Very High EPEAK was Demonstrated
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
23
Cryomodule Assembly
2009, OPT=0.15
P.N. Ostroumov, Warm and Cold Ion Linacs
2013, OPT=0.077
Collaboration Meeting
March 30, 2015
24
Design evolution of the injector linac: Cost
Reduction
 Number and type of SC cavities is reduced since 2010 design
 Footprint is reduced significantly
 Based on SRF technology development for ATLAS facility
EPEAK ,
MV/m
BPEAK,
mT
Cavity types
# of
Cavities/Cryos
Linac length,
m
2011
30
80
QWR, HWR, Spoke
119/19
150
2012
30/40
90
QWR,HWR
122/16
121
2014
50
80
QWR,HWR
109/13
~110
2015
60
80
QWR,HWR
96/12
~100
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
25
Lorentz Detuning
 Lorentz force detuning coefficient
– 72 MHz QWR: -1.6 Hz/(MV/m)2
– 170 MHz HWR: -1.5 Hz/(MV/m)2
 Simplest method is to provide high RF power in overcoupled mode
like SNS
– Expensive way, high power is not required for EIC injector
 Piezoelectric tuner is a common way to compensate for Lorentz
detuning in pulsed SC cavities
 There is also issue for LLRF to provide stable phase-locked
operation
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
26
Summary
 SC Linac is the best option for acceleration of multi-ion beams
– Protons – 280 MeV
– Lead ions -100 MeV/u
 Performance of SC cavities is being improved every year
– Cost of a SC linac is stable in “cost of living” dollars
 Multi-gap NC Linac for 100 MeV/u lead ions requires 417 MV
effective voltage
– Protons can reach to 100 MeV only
 NC linac composed from 2-gap (or 3-gap) cavities requires large
cavity count, ~500 cavities
– This results in very high cost of both cavities and RF system
P.N. Ostroumov, Warm and Cold Ion Linacs
Collaboration Meeting
March 30, 2015
27